def kuwahara(arr, size=5):
"""
Kuwahara, a nonlinear 2D smoothing filter.
http://subsurfwiki.org/wiki/Kuwahara_filter
Args:
arr (ndarray): a 2D array, such as a seismic horizon.
size (int): the kernel size, e.g. 5 for 5x5. Should be odd,
rounded up if not.
Returns:
ndarray: the resulting smoothed array.
"""
def func(this, s, k):
"""
Deal with this patch.
"""
t = this.reshape((s, s))
sub = np.array([t[:k, :k].flatten(),
t[:k, k-1:].flatten(),
t[k-1:, :k].flatten(),
t[k-1:, k-1:].flatten()]
)
select = sub[np.argmin(np.var(sub, axis=1))]
return np.mean(select)
arr = np.array(arr, dtype=np.float)
if arr.ndim != 2:
raise BrugesError("arr must have 2-dimensions")
if not size // 2:
size += 1
k = int(np.ceil(size / 2))
return scipy.ndimage.generic_filter(arr,
func,
size=size,
extra_keywords={'s': size,
'k': k,
}
)
def snn(arr, size=5, include=True):
"""
Symmetric nearest neighbour, a nonlinear 2D smoothing filter.
http://subsurfwiki.org/wiki/Symmetric_nearest_neighbour_filter
Args:
arr (ndarray): a 2D array, such as a seismic horizon.
size (int): the kernel size, e.g. 5 for 5x5. Should be odd,
rounded up if not.
include (bool): whether to include the central pixel itself.
Returns:
ndarray: the resulting smoothed array.
"""
def func(this, pairs, include):
"""
Deal with this patch.
"""
centre = this[this.size // 2]
select = [nearest(this[p], centre) for p in pairs]
if include:
select += [centre]
return np.mean(select)
arr = np.array(arr, dtype=np.float)
if arr.ndim != 2:
raise BrugesError("arr must have 2-dimensions")
if not size // 2:
size += 1
pairs = [[i, size**2 - 1 - i] for i in range(size**2 // 2)]
return scipy.ndimage.generic_filter(arr,
func,
size=size,
extra_keywords={
'pairs': pairs,
'include': include
})